Clinical RESEARCH/AUDIT

Diagnosing wound infection:

the use of C-reactive protein
Andrew Kingsley and Vanessa Jones

Abstract
Background: The diagnosis of wound infection is often restricted to the use of clinical signs and symptoms which are not always
reliable. Aims: To determine if systemic C-reactive protein (CRP) measurements could be used as a diagnostic marker for wounds
whose healing is delayed by the effects of wound bed bioburden. Methods: A comparative descriptive design with a survey method
was used. Patients with wounds healing by secondary intention with a duration of 4 weeks or more were placed into wound
infection continuum groups based on clinical features.Wound features for each patient were listed to enable analysis of CRP
results within and between groups. CRP was also analysed for all patients as a single group against individual wound features.
Results: CRP levels were found to be significantly higher for patients in the spreading infection group, but no differences in levels
were found to distinguish the other groups. Higher CRP levels were found to be statistically significant when several clinical
features were present in wounds irrespective of study grouping, these were necrotic tissue, wet wounds, malodour, recent wound
extension and redness of >2cm on surrounding skin. Conclusions: This study cannot support CRP as a diagnostic marker for
determining between wounds in different groups in the infection continuum known as colonisation, critical colonisation and local
infection and cannot be used alone to indicate the need for antimicrobial treatment in these groups. Conflict of interest: None

antibiotic use for patients with chronic
wounds is higher than for other patients
KEY WORDS matched for age and sex. This raises costs
C-reactive protein and contributes to the development
Wound infection continuum and selection of multi-resistant micro-
Secondary intention organisms (Howell-Jones et al, 2005).
Critical colonisation
Diagnosis of infection in wounds
Local infection healing by secondary intention is
primarily undertaken through the
interpretation of clinical signs and
symptoms. Gardner et al (2001a)
When Cutting (1998) compared
surgical nurses’ clinical diagnoses of
infection and his own judgements
based on a similar list (Cutting and
Harding, 1994) he found a poor
agreement between the author and
the nurses for both infected and not
infected wounds (47.5%), indicating
that clinical signs and symptoms
alone are not sufficient to reliably
identify infection.

M
any wounds healing by
secondary intention become
static causing extended periods
of discomfort and inconvenience for the
patient, and an increase to healthcare
costs and workload. There are many
reasons for this delay and the effect of
wound bioburden is a common one
(Browne et al, 2001) although its exact
role is uncertain (Howell-Jones et al,
2005; Penhallow, 2005). The suggested
range of the prevalence of chronic
wound infection lies between 0.14%
and 6% (Verdu Soriano et al, 2004) and
Andrew Kingsley is Clinical Manager Infection Control
and Tissue Viability, Northern Devon Healthcare Trust,
and Vanessa Jones is former Senior Lecturer/
Education Director (now retired), Wound Healing
Research Unit, University of Wales
developed a checklist tool from these Qualitative microbiology provides
based on an earlier list (Cutting and secondary corroboration. However,
Harding, 1994). Known as the Clinical the link between positive wound
Signs and Symptoms Checklist, it swabs and clinical cases of infection
identifies 12 signs and symptoms of has been shown to vary in wounds
infection which are: with different aetiologies. Schmidt
8 Pain et al (2000) found that only 22% of
8 Erythema venous leg ulcers (VLU) with positive
8 Oedema swabs went on to develop infection
8 Heat in contrast with 70% of diabetic and
8 Purulent exudate arterial ulcers. McGuckin et al (2003)
8 Serous exudate with concurrent considered that microbiological
inflammation culturing of wounds was insufficiently
8 Delayed healing rigorous as a scientific method,
8 Discolouration of granulation requiring further research to
tissue improve its validity as a predictor
8 Friable granulation tissue of wound infection.
8 Pocketing at the base
8 Foul odour Biopsy is considered the gold
8 Wound breakdown. standard microbiological method for

32 Wounds UK, 2008, Vol 4, No 4

p32-46Kingsleyok.indd 2 29/10/08 00:15:22


research but is not often carried out
in clinical practice due to exacting
processing requirements and concerns
over harm to the patient and so
swabbing is more usual. Quantitative
thresholds as indicators of infection
have been used for many years
(Cooper, 2002a), although the value
for this method for open wounds
has been questioned (Bowler, 2003).
Wound surfaces and tissues have
heterogeneous bacterial distribution,
which presents a problem for testing
accuracy (Woolfrey et al, 1981;
Steer et al, 1996; Ribinik et al, 1997).
Problems with swabbing consistency
between professionals may lead to
unreliable results (Cooper, 2002a). In
one study, Starr and MacLoed (2003)
found varying techniques in wound
swabbing in a small group of nurses
from the same ward. As there is no
UK consensus on swabbing method in
clinical practice including the value of
pre-cleaning the wound bed (Cooper,
2002b), it is likely that results of
swabbing will lack consistency.
It has been proposed that a
continuum of gradually-increasing
bioburden exists in open wounds
ranging from healing wounds with
simple colonisation and without host
response to overt infection (Dow et
al, 1999). Names for the staged points
on this infection continuum have been
suggested (Kingsley, 2001; Kingsley
et al, 2005) and are colonisation,
critical colonisation, local infection,
and spreading infection. The term
critical colonisation, first introduced
by Davis (1997), has been the subject
of debate (Cooper, 2005; Jorgensen et
al, 2005; White and Cutting, 2006) but
is used by many wound care experts
to identify when topical antimicrobial
therapy is required to optimise
wound healing (Jorgensen et al, 2005).
However, clinical diagnosis, being
dependent on the skills of the clinician,
leaves scope for either over-treatment
with antimicrobials when it is not truly
warranted or under-treatment when
managing the bioburden would assist
wound healing.
If a simple and reproducible
diagnostic test was available to indicate
p32-46Kingsleyok.indd 3
Clinical RESEARCH/AUDIT
Table 1
Characteristics of wounds in each of the wound infection continuum group
Colonised
A wound that is considered to be progressing normally. There will be no necrotic tissue, and limited
slough which is thin and mobile in depth and character. There will be some firm granulation tissue
which is normally a salmon red-pink colour and moist, and there may be pink/purple epithelial tissue
(Sibbald et al, 2000).
Critically colonised
A wound that is considered to be indolent (ie not healing and not deteriorating) despite seemingly
appropriate therapy. The wound is wet and there is likely to be thick slough (at >50% cover of base)
but no new (within two weeks of referral to tissue viability) necrotic tissue (there may be residual ne-
crosis from original wounding event), there may be odour but not cellulitis. The wound may bleed easily
at slightest touch and if previously granulating the tissue may be an unhealthy dark colour. Alternatively
the wound base may be clean, free of active granulation and pale in colour or pale and oedematous.
Locally infected
A wound that has some cellulitis as a flare on the adjacent skin or encircling the wound no greater
than 2cm (Dow et al, 1999) from the wound edge. There may be slough (>50% of base) and necrotic
tissue (appeared suddenly in a previously healing wound and this may be in the absence of cellulitis
(Kingsley, 2001), the wound base if previously granulating may be an unhealthy deep red colour that
bleeds at the slightest touch, the wound is wet or wetter than it was recently, and there may be
odour. Some wounds may not exhibit cellulitis but show recent wound extension. Others may show
cellulitis <2cm from the edge but show extension.
Spreading infection
There is cellulitis surrounding the wound that is greater than 2cm from the edge. The wound may
be wetter than normal, tissue may be friable and bleed easily, the wound may have extended or new
satellite skin breaks have appeared in addition to the cellulitis. Odour may be offensive, there may be
pyrexia, lymphangitis or lympadenitis or bacteraemia.
when the use of antimicrobial therapy of classic signs of infection, could be
would be necessary for wounds that undertaken with confidence. Equally
were not healing due to the effects if high CRP levels were found to
of wound bed bioburden then early associate with particular wound
treatment might improve outcomes features or certain micro-organisms
and lower costs. Measuring C-reactive then this could form the basis for
protein (CRP) levels is a potential further research into the prediction of
test that has been used successfully infection.
in other areas to improve diagnosis
over standard bacteriology and clinical CRP testing — a possible solution
impression (Gulich et al, 1999), such CRP is a member of the acute
as identifying sepsis in major trauma phase family of proteins, released
(Stahl et al, 1985), and detecting from hepatocytes whose levels rise
wound infection following trauma and significantly in response to monocyte-
orthopaedic surgery (Gupta et al, origin mediators such as interleukin-
2002). If it could be established that 1 (IL-1) and 6 (IL-6), because of
CRP levels distinguish between the infection, tissue injury (Houshian
different wound infection continuum et al, 2000) and inflammatory or
states then prompt initiation of neoplastic diseases (D’Amore, 2005).
treatment of the abnormal states, CRP is a highly conserved protein
particularly critically colonised wounds and no deficiency has been noted
with delayed healing and absence in any human population studied
Wounds UK, 2008, Vol 4, No 4 33
29/10/08 00:15:22
 Clinical RESEARCH/AUDIT
(Ahlers and Schonheyder, 1990; Du
Clos and Mold, 2002). Changes in
CRP levels can occur within 24–48
hours in response to the presence
or removal of the stimulation (Pepys,
1981), such as a bacterial infection.
CRP elevates rapidly in response to
relevant stimulus and is considered to
have a stable decay rate (Okamura et
al, 1990). Larsson et al (1992) found
that the post-surgery CRP response
was not affected by use of prophylactic
antibiotics or anti-inflammatory drugs
indicating that it reacts to a relevant
stimulus in the body and is not
altered by circulating medications. It
is, therefore, reasonable to presume
that CRP levels, elevated because
of infection, will fall in response to
effective antimicrobial therapy. CRP is
a non-specific marker of inflammation
and its potential as a diagnostic
indicator for wound infection may be
restricted depending on the presence
of other inflammatory pathologies in
individual patients. However, vascular
vessel damage associated with venous
disease was found not to raise CRP
above routine clinical levels (Blomgren
et al, 2001) and Goodfield (1988)
found that venous eczema was also
unable to raise CRP levels.
There is currently no single
predictor of wound infection that
is considered perfect (Browne et al,
2001). This study aims to investigate
the possibility of using CRP levels to
categorise infection levels.
Method
A comparative descriptive design
using a survey method was used
to search for a link between CRP
level and open wounds categorised
into the four states of the wound
infection continuum (colonisation,
critical colonisation, local infection and
spreading infection) (Table 1). Ethical
approval was received from the North
and East Devon Local Research Ethics
Committee. Sixty-four patients with
wounds of more than four-weeks
duration were recruited purposively
from a single UK tissue viability nursing
service. Exclusion criteria, based on
work by Goodfield (1988), were
stipulated to prevent inclusion of
34 Wounds UK, 2008, Vol 4, No 4
p32-46Kingsleyok.indd 4
Table 2
Essential and corroborating criteria used to allocate wounds into study groups
Allocation group Essential criteria (must
exhibit at least one)
Colonised (C) 1. Progressing normally
2. No necrotic material
3. Some granulation
Critically colonised 1. Is indolent
(CC) 2. Is wet
3. No new necrosis
4. No cellulitis
Locally infected 1.Cellulitis — as flare or en-
(LI) circling wound — no greater
than 2cm
2.If no cellulitis wound has
extended recently
3.If no cellulitis sudden
appearance necrotic tissue
4.Is wet or wetter than it
has been
Spreading infection 1.Cellulitis >2cm surrounding 1. May have offensive odour
(SI) the wound
patients who might have had reasons
other than wound infection for a CRP
rise, fall or suppression that would
reduce the validity of the results.
Exclusions were acute trauma wounds
(<4 weeks), surgery or a myocardial
infarction within the preceding
seven days, an unresolved infection
elsewhere in the body, and patients
with active inflammatory conditions
such as rheumatoid arthritis,
inflammatory bowel disease, vasculitis,
pyoderma gangrenosum or Buerger’s
disease. Patients with diabetes and
foot ulceration were excluded because
the associated neurological damage
and microvascular changes can lead
to masking of the usual visual cues or
experienced symptoms for infection,
notably through absence of pain and
inability to vasodilate in response to
infection (Edmonds and Foster, 2004;
Corroborating criteria (may exhibit 0–all)
1. Granulation tissue normally
red/pink and moist
2. There may be pink/purple epithelium
1. Likely to be thick slough (>50% cover of base)
2. Possible odour
3. May be friable
4. May have unhealthy dark granulation
5. Granulation may be absent
6. Base may be pale or pale and oedematous
1. May have thick slough (>50% cover of base)
2. May have odour
3. May be friable
4. May have unhealthy dark granulation
5. Granulation may be absent
6. The base may be pale or pale and oedematous
2. May be friable
3. May be wetter than normal
4. May have wound extension
5. May have new satellite lesions
6. May have pyrexia, lymphangitis, lymphadenitis
or bacteraemia
Falanga, 2005). Patients were also
excluded if they had conditions or
were taking medications that might
suppress the normal inflammatory
responses, such as systemic
steroids, cancer chemotherapy, and
antimicrobials taken or applied for
greater than 48 hours, as they
are either direct or indirect
inflammation suppressants.
Eligible subjects were allocated
to one of the four wound infection
continuum study groups following
comparison of clinical signs and
symptoms to a pre-defined checklist
based on published expert views
available before the start of data
collection in 2003 (Cutting 1994;
Cutting and Harding 1994; Thomson
and Smith 1994; Davis 1997; Davis
1998; Dow et al, 1999; Sibbald et
29/10/08 00:15:23
 Clinical RESEARCH/AUDIT

al 2000; Kingsley 2001) (Table 2).

Table 3 As a secondary aim of the study
Linkage of clinical feature to a mean raised CRP level (>8mg/L) as tested by ANOVA microbiological samples were taken
to look for any potential associations
between the organisms identified,
Statistically significant Not statisitically significant the clinical groups of the wound
Necrotic tissue (p < 0.001) Wound type (not ANOVA tested) infection continuum and wound
features such as necrosis and CRP
Wet or wetter than normal (defined by patient Diabetes (p = 0.306)
level. Following group allocation by the
or by researcher as obviously wet on observation)
(p=0.017) single researcher swabs for aerobic
and anaerobic analysis were taken
Malodour (p=0.026) Granulation (healthy and unhealthy including using a standard procedure defined as
overgranulation and discoloured granulation) (p protocol 2 by Kingsley and Winfield-
= 0.105)
Davies (2003). Venous blood samples
Recent wound extension (p=0.01) Friable/bleeding tissue (p=0.503) for CRP assay were collected by
Redness (>2cm from wound edge) (p=<0.0010) New ‘satellite’ lesions (p=0.586) vacuum specimen tube (Vacuette®,
Spreading infection group (p=<0.001) Redness (<2cm from wound edge) (p=0.344) Greiner Bio-One Ltd, Gloucestershire).
Wound size was measured with single-
use, clean, non-sterile paper metric
rulers using the perpendicular method
Table 4 to estimate surface area after the
technique used by Daltrey et al (1981).
Further results on clinical features related to the allocated wound infection continuum study groups
Clinical features were recorded
for each of the 64 wounds so it was
Clinical feature Result possible to analyse CRP levels against
Tissue discolouration 8 Discolouration present in 30% of CC group and 16% LI group wound features as a single series
8 Darkly discoloured wounds had lower mean CRP (36.8mg/L) compared irrespective of the clinical group to
with non-discoloured wounds (50 mgl/L) which the patient had been allocated
Friable tissue 8 Found most commonly in LI group (4 of 13 wounds) for the main aim of the study. The
8 None of the four LI group wounds with friable tissue had concomitant following clinical features were
cellulitis examined against CRP levels:
8 Wound type
Overgranulation 8 Overgranulation was only found in the CC group (4 of 20 wounds)
8 Diabetic status
and LI groups (2 of 13)
8 Four of the six (66.6%) overgranulated wounds bled easily compared 8 Wound size
with 3 of the 58 (5.2%) without overgranulation present 8 Wound age
8 Granulation tissue
Necrotic tissue 8 Five of 7 wounds which still had necrotic tissue present from the 8 Discoloured tissue
original injury were categorised as infected (LI and SI)
8 Friable/bleeding granulation tissue
8 Wounds containing necrotic tissue had a higher mean CRP than
those without it (p=<0.001) 8 Overgranulation
8 Necrotic tissue
Wet (researcher These wounds had a higher mean CRP than those wounds considered to have 8 Sloughy tissue
defined) or wetter lower levels of exudation, irrespective of clinical group allocation (p=0.017) 8 Exudate level
than normal (patient
8 Malodour
defined) wounds
8 Wound extension (extension
Malodour 8 Malodour was part of the CC group descriptors but was actually found of margins and those with new
most commonly in the LI group (5 of 13 wounds) ‘satellite’ lesions beyond the
8 High mean CRP and malodour were significantly linked (p=0.026) immediate margins of the pre-
Wound extension Recent extension of the wound had occurred in 11 of the 64 study wounds existing wound)
and was linked to a high mean CRP (p=0.01). Half (n=8) of the wounds with 8 Cellulitis (<2cm and >2cm from
four bacterial species isolated had extended, as had 15.3% of wounds with the wound edge).
three isolates, 17.6% with two isolates and 30% with 1 isolate, and 20% of
wounds recording ‘no growth’.
Results
Indolence (delayed Mixed bacterial cultures coincided more frequently in CC wounds which were The design was non-experimental
healing) indolent by study definition. Only 1 of 10 wounds with ‘no growth’ were and patient sub-groups were analysed
allocated to the CC group, and in contrast 7 of 17 with 2 isolates, 7 of 13 descriptively using summary statistics.
with 3 isolates and 2 of 4 with 4 isolates were allocated to the CC group Statistical analysis of CRP levels
against wound groups and features

36 Wounds UK, 2008, Vol 4, No 4

p32-46Kingsleyok.indd 6 29/10/08 00:15:24

 Clinical RESEARCH/AUDIT
was undertaken using a one-way
unrelated ANOVA with the software
SPSS version 11.5. The total 64 cases
were allocated using clinical signs and
symptoms into the four different study
groups, 21 into colonised (C), 20 into
critically colonised (CC), 13 into local
infection (LI), and 10 into spreading
infection (SI). CRP data was missing
on three cases. Two were in the LI
and one in the SI groups. Analysis of
CRP results and their linkage to clinical
features within and between groups
were undertaken in 61 cases, whereas
simple percentage calculations
on the presence or absence of
clinical features were based on the
full 64 cases.
A clear difference between the SI
group and the other groups emerged
when the means of the CRP levels
were compared (ANOVA p=<0.001).
The SI group mean was 167.88mg/L
(standard deviation (SD)=122.17).
All of the other groups had mean
CRP values that were above normal
(equal to or less than 8mg/L): C=27.33
(SD=27.87), CC=33.85 (SD=32.22),
LI=29.54 (SD=32.94). However, it was
not possible in this study to distinguish
with statistical significance between
the C, CC and LI groups using CRP as
the diagnostic marker.
For wounds with an extending
cellulitis (SI) a statistically significant
higher mean CRP was recorded
against all the other groups in
the study.
Wound types in the study were
mostly venous ulcers (n=26; 40.6%),
with the remainder being made up
of pressure ulcers (n=9), ar terial
(n=1), mixed (n=8) and hydrostatic
leg ulcers (n=3), trauma (n=4),
surgical wounds (n=7), hear t failure-
associated leg-swelling ulcerations
(n=2), friction rub from or thotic shoe
(n=1), oedema blister (n=1), wound
over an old poliomyelitis operation
site (n=1) and a wound of unknown
aetiology (n=1).
Wound extension
Extension occurred in 17.18% of the
wounds in this study suggesting it to
38 Wounds UK, 2008, Vol 4, No 4
p32-46Kingsleyok.indd 8
be a feature worthy of consideration.
Those wounds with recent extension
had a statistically significant higher
mean CRP than those without.
Microbiology results
In this study 16 wounds contained an
isolate of Streptococcus either group
B (n=3), D (n=8) or G (5). Extension
was recorded on five occasions, of
which group G was the most common
(n=3), with B and D both having a
single wound extension.
Wound duration
The relationship between CRP levels
and wound age was tested and was
not found to be significant (p=0.156).
The SI group had the lowest mean
age (3.22 months) against a combined
mean of all the other groups (C,
CC and LI) (20.27 months), but this
did not produce a significant result
(p=0.232).
Trauma
The four wounds originating from
trauma at four weeks or greater from
enrolment to the study all featured in
the problem groups, with one in the
CC and three in the SI group.
Discolouration
Discolouration was found in 10
wounds and was more common in
CC than LI wounds. In the 20 CC
wounds, six were discoloured and five
of these were darkly so. These dark
wounds had a higher mean CRP result
than those in the CC group without
this dark discolouration. However,
when comparing mean CRP in all the
darkly discoloured wounds across the
categories with all the wounds without
discolouration, the mean CRP was
lower. The most common organisms
found in the darkly discoloured CC
cases were coliforms (4/5). Of the
total seven wounds in all the study
groups with dark discolouration five
were VLU and one was a mixed
venous/arterial ulcer.
Necrotic tissue
The presence of raised CRP levels
and necrotic tissue did coincide,
though statistical significance may have
been skewed as 5 of the 11 wounds
containing necrotic tissue were in the
SI group. However, the two wounds
with necrotic tissue present in the
colonised group also had elevated
CRP levels (16mg/L and 54mg/L)
suggesting that necrotic tissue is
an impetus to inflammation. Given
that two (15.38%) of the total 13 LI
cases had necrosis remaining from
the outset of the wound, as did 3/10
(30%) SI cases, it would suggest that
necrosis from the original wounding is
a risk factor for infection. This finding
concurs with the results from Gardner
et al (2001a) and further supports
the case for debridement to improve
healing made by Steed et al (1996)
and Williams et al (2005) by reducing
likelihood of delay through the
occurrence of infection. Taking all the
seven wounds with residual necrosis
from original injury it was found that
five (71.42%) had become infected by
the time of study enrolment. Therefore
original residual necrosis was linked
to outcome of infection, and the
presence of necrosis (original or new)
to higher CRP.
Exudate levels
Wounds with exudate levels that were
considered wet or wetter than normal
had a significantly higher mean CRP.
Odour
Ten of the 64 (15.62%) wounds in
this study were malodorous, as defined
subjectively by the researcher.
There was a statistically significant
link between malodour and a high
mean CRP.
Discussion
The purpose of the study was to
determine if different CRP levels were
associated with wounds in different
clinical states of the wound infection
continuum. If clear differences between
the levels of CRP were found to exist
for each of the clinical groups of the
wound infection continuum then a
potential diagnostic test will have
been identified for further research.
In addition an aggregate view of the
relationship of CRP with individual
clinical features and the risk of
infection were explored and several
interesting findings were made.
29/10/08 00:15:25
 Clinical RESEARCH/AUDIT
The total number of infected cases
in the study (LI and SI) defined by
clinical features was 23/64 (35.93%).
The sample reflected the normal
clinical situation so infection appears
to be common in open wounds older
than four weeks referred to the tissue
viability service in North Devon. If all
wounds considered clinically to have
bioburden problems are included (CC,
LI, SI) this figure becomes 67.18%
(43/64). If this pattern is repeated
across the UK then it would seem
paramount that tissue viability nurses
have a good understanding of the
management of CC and infection.
Wound extension
Extension might result from increased
wetness macerating the wound edge
or an active invasion of bacteria into
wound edge tissues, although both
would be expected to elicit a new
or heightened inflammatory reaction
and hence a rise in CRP. Extension
did occur in 11 of the 64 wounds
(17.18%) in this study and CRP level
elevation was statistically significantly
greater than for wounds without
extension. New satellite lesions were
less common than expansion of the
circumference and CRP levels for
those wounds with and without this
feature did not differ significantly.
Microbiology results
Various studies have made links to
specific organisms being the cause
of healing delay, extension and larger
ulcer size, for example Hansson et al
(1985) point to the fungus Candida
albicans; Schraibman (1987) identified
beta haemolytic streptococci; Halbert
et al (1992) reported anaerobes;
Hansson et al (1995) identified
Pseudomonas aeruginosa, and
Madsen et al (1996) P aeruginosa,
Staphylococcus aureus and beta
haemolytic streptococci (groups A, B,
C and G).
Schraibman (1987) reported
a frequency of beta haemolytic
streptococci at 30% similar to the
percentage in this study of 64 mixed
wound types at 25% (16 wounds).
Wound extension in the presence
of beta haemolytic streptococci
40 Wounds UK, 2008, Vol 4, No 4
p32-46Kingsleyok.indd 10
occurred only in the infected groups
demonstrating their accepted
propensity to invasive infection and
ability to increase wound size. The
infection rate, described by clinical
signs and symptoms of LI and SI in this
study, for beta haemolytic streptococci
was 50% (8/16). However, on none
of these eight occasions were the
streptococci cultured alone, being
found together with anaerobes (n=2),
Proteus (n=2), S. epidermidis (n=2), S.
aureus, methicillin-resistant S. aureus,
coliforms, and diphtheroids (all n=1).
On the only occasion when a beta
haemolytic streptococcus was cultured
alone it was found in a wet but static
CC wound, rather than in a wound
displaying any of the classic signs of
infection or cellulitis. Examination
of the number of isolates and the
presence of extension showed that
wounds yielding four isolates were
more likely to extend (50%). Trengove
et al (1996) reported that four or
more isolates caused an increased
chance of failure to heal in leg ulcers
so the finding that 50% of wounds
with four isolates in this study were
actively extending seems in keeping
with their findings. The purpose of the
study was not to link the microbiology
of the study wounds to wound size
so the results reported here are
observations only.
Trauma
The aetiology of trauma did appear to
influence the subsequent development
of spreading infection after four weeks,
with three of four (75%) wounds of
this type presenting with this problem.
While numbers in the study were too
small to make any definite conclusion
it would seem that trauma may be
linked to greater risk of acquiring a
spreading infection than for other
wound types. One possibility for this
is that traumatic injury may inoculate
pathogens deeper into the tissues
evading skin level defence mechanisms
aimed at localising infection.
Discolouration
It has been identified that clinicians
use discolouration as an aid to
diagnose infection in open wounds
(Cutting, 1998; Bamberg et al, 2002;
Cutting et al, 2005) even though
it has been reported to have low
validity for this purpose (Gardner
et al, 2001b). Cutting et al (2005)
found that clinicians considered dark
discolouration to be a sign of infection
in VLU. In this study 10 wounds
had discolouration of which seven
were subjectively assessed as being
darkly discoloured. Five of the seven
darkly discoloured wounds were
from the CC group, one from the LI
and one from SI groups. The mean
CRP value was lower in the darkly
discoloured wounds in comparison
with the 51 other wounds without
any discolouration but there was no
statistical significance to this finding.
The single case of dark discolouration
in the SI group had a CRP of 12mg/L,
while the other eight cases in the SI
group all had normal tissue colour and
all but one (CRP = 10mg/L) had much
higher CRP levels (range 51–339mg/L).
Necrotic tissue
Seven wounds had necrotic tissue
present remaining from the original
injury. There were 2 (9.52%) in the
C group (n=21), 2 (15.38%) in the LI
group (n=13) and 3 (30%) in the SI
group (n=10). Given that 5 (71.42%)
of 7 wounds with necrotic tissue
remaining from the original injury
were classified as having either local
or spreading infection, this finding is
consistent with the accepted view
that necrotic tissue is a risk factor
for infection. A total of 11 wounds
had necrotic tissue present either
remaining from original injury or
more recently formed, and these
wounds had a higher mean CRP
(p=<0.001) than those wounds
without necrotic tissue. This
finding concurs with the results
from Gardner et al (2001a) and
fur ther suppor ts the case for
debridement to be carried out to
improve healing by reducing the
likelihood of delay through the
occurrence of infection (Steed et al,
1996; Williams et al, 2005).
Exudate levels
Wounds with more exudate than
normal had a significantly higher
mean CRP level. As wetness is a sign
29/10/08 00:15:26
 Clinical RESEARCH/AUDIT
of inflammation (Thomas, 1997) this
finding concurs with expectations.
Interestingly, wetter LI wounds had
a distinctly higher mean CRP than
their drier LI counterpar ts. Thus even
within a single clinical group where all
the wounds were diagnosed clinically
as infected, increased wetness
demonstrated those mounting a
greater inflammatory response. This
could indicate that the situation
is more severe in cer tain cases or
that host response is more dynamic
in different individuals. A limitation
to the use of subjectively assessed
exudate levels is that wetness
may be a product of poor dressing
choice or infrequent dressing change
rather than a physiological response
to infection.
Odour
Odour is reasonably common with
a 9% prevalence (Lindholm et al,
2005), which is greater (36%) in
chronic ulcers at risk of local infection
(Meaume et al, 2005) and critically
colonised wounds (55%) (Jorgensen
et al, 2005), showing the link between
malodour and delayed healing.
Antimicrobial strategies reduce odour
(Kalinski et al, 2005) and restart
healing (Jorgensen et al, 2005) linking
odour to micro-organisms.
Lindholm et al (2005) surveyed
all wounds including closed surgical
wounds that would not be expected
to produce malodour, so with this in
mind the current study found fewer
malodorous wounds than expected.
Micro-organisms considered to be
capable of producing odour are Gram-
negative bacilli, proteus and anaerobes
(Cutting and Harding, 1994; Bowler
et al, 1999). Anaerobes were present
in only 4/10 malodorous wounds.
Fourteen wounds contained anaerobes
in all of which four (28.57%) had
malodour, which is consistent with the
rate reported by Dow (2001), which
was stated as less than 50%. Proteus
species were present in eight wounds
of which three (37.5%) had malodour,
and 3/17 (17.64%) wounds with
coliforms were odorous. The strongest
coincidental isolate to odour was
therefore Proteus.
42 Wounds UK, 2008, Vol 4, No 4
p32-46Kingsleyok.indd 12
The 10 malodorous wounds in
this study also had exudate levels that
were ‘wet or wetter than normal’.
Although CRP was elevated in all
study groups, malodour was found in
association with significantly higher
CRP, suggesting that the organisms
responsible for odour generation were
also capable of inducing a greater
inflammatory response as identified
by raised CRP levels, hence the
wetness, and in some cases invasion
with cellulitic response. Malodour,
like exudate levels, can be linked to
dressing choice and the frequency of
dressing changes, and this along with
the subjectivity required to assess it,
limits the results.
CRP levels as indicators of infection
Cellulitis is the expression of the
inflammatory response mediated
directly by the invading organisms
and indirectly by the immune system
attempting to eradicate those
organisms and is a key sign of infection
(Eron et al, 2003). In this study there
was less cellulitis in the LI group than
was expected by the authors. Cellulitis
at <2cm from wound edge (which
was one of the diagnostic features
for the LI group) was not associated
with a statistically significant rise in
CRP, whereas wounds with recent
extension of surface area did have
higher mean CRP compared with
wounds without that feature.
The SI group showed a statistically
significant higher mean CRP against
all the other cases in this study. This is
clearly a group that needed treatment
with systemic antibiotics as advocated
by Eron et al (2003) in order to
prevent potential systemic morbidity
from the wound infection.
Conclusion
As the SI group was the only group
that showed a significantly higher
mean CRP, and all the other groups
were indistinguishable by use of CRP,
this study was unable to provide
any evidence that CRP may be a
useful marker for determining the
clinical states of colonisation, critical
colonisation and local infection. The
finding that LI wounds did not having
any significant CRP elevation over
the CC group, or indeed over the
C group, suggests that a potential
therapeutic rationale for topical
antimicrobial treatment to deal with
this problem based on clinical signs
and symptoms could be considered.
Regular CRP measurement could
therefore be used to monitor
whether it is necessary to move to
systemic antibiotics in the event of a
sudden and definite rise. This study
did find that wounds older than four-
weeks duration healing by secondary
intention irrespective of clinical signs
and symptoms of infection commonly
had raised CRP, indicating an active
inflammatory host response.
Mean CRP values, gathered
on 61 of the 64 cases, were
statistically significantly higher for a
number of wound features in this
study when all 64 cases were analysed
as a single group. These features
were redness >2cm from wound
edge, necrotic tissue, malodour, wet
or wetter than normal wounds, and
wound extension.
Limitations and implications for future
study
A number of significant limitations
must be noted when considering the
data produced in this study. Survey
design is about describing only what
things are like and not why they occur
like that. Surveys are, therefore, useful
in the development of hypotheses for
deductive testing to determine cause
and effect relationships. A convenience
sample was used so unlike a random
sample it has a greater risk of bias.
The work was undertaken by a single
researcher who made the allocations
to the wound infection continuum
study groups using an unvalidated
tool. Exclusion criteria did not include
patients taking non-steroidal anti-
inflammatory drugs which may have
had the potential to alter CRP levels
and are drugs in common use among
older people who were strongly
represented in the sample. Clinical
features, such as odour and exudate
levels were assessed without recourse
to a validated tool, so internal
reliability was compromised. The small
29/10/08 00:15:27
 Clinical RESEARCH/AUDIT
numbers of patients in each of the
study groups were insufficient
to determine anything other than
large and very obvious differences in
CRP measurements.
Future studies could be directed
at identifying which clinical features
are indicative of the state of
bioburden-influenced delayed healing
without visible inflammatory host
response (critical colonisation), and
then validating a tool based on that
information. Such a tool could then
be used to identify wounds requiring
antimicrobial interventions in clinical
practice, or used as a base to recruit
individuals into new research projects
on infection in open wounds.
Finding a simple, reliable and
rapid method of diagnosing wounds
experiencing the detrimental effects of
the bioburden they support remains
an important goal. Should such a test
be found it would facilitate a change
in the way we approach non-healing
wounds, reducing reliance on clinical
diagnosis of bacterial interference in
the healing process that is open to
both over- and under-recognition.
Early intervention with antiseptic or
new novel therapies could reduce
the number of wound infections
that cause added discomfort for the
patient, lengthen the duration to
healing, require antibiotics and elevate
healthcare costs. WUK
This study was undertaken as part of
a course for the award of MSc Wound
Healing and Tissue Repair, Cardiff
University.
References
Ahlers A, Schonheyder H (1990) C-reactive
protein in patients with infection. Ugeskr
Laeger 153(1): 13–16 (English abstract on
Medline)
Bamberg R, Sullivan P, Conner-Kerr T
(2002) Diagnosis of wound infections:
current culturing practices of US wound
care professionals. Wounds 14(9): 314–27
Blomgren L, Johansson G, Siegbahn A,
Bergqvist D (2001) Coagulation and
fibrinolysis in chronic venous insufficiency.
Vasa 3: 184–7
44 Wounds UK, 2008, Vol 4, No 4
p32-46Kingsleyok.indd 14
Bowler P (2003) The 105 bacterial growth
guideline: Reassessing its clinical relevance
in wound healing. Ostomy Wound Manage
49(1): 44–53
Bowler P, Davies B, Jones S (1999)
Microbial involvement in chronic wound
malodour. J Wound Care 8(5): 216–18
Browne A, Dow G, Sibbald R (2001)
Infected wounds: definitions and
controversies. In: Falanga V, ed. Cutaneous
Wound Healing. Martin Dunitz Ltd, London
Cooper R (2002a) The contribution of
microbial virulence to wound infection. Br
J Nurs 11(22)Suppl: 10–14
Cooper R (2002b) Wound microbiology:
past, present and future. Br J Nurs 11(22)
Suppl: 4–6
Cooper R (2005) Understanding wound
infection. In: European Wound Management
Association (EWMA). Position Document:
Identifying Criteria for Wound Infection.
MEP Ltd, London
Cutting K (1998) Wounds and Infection.
Wound Care Society (educational booklet),
London
Cutting K (1994) Detecting infection.
Nursing Times 90 (50): 602
Cutting K, Harding K (1994) Criteria for
identifying wound infection. J Wound Care
3(4): 198–201
Cutting K, White R, Mahoney P, Harding
K (2005) Clinical identification of wound
infection: a Delphi approach. In: European
Wound Management Association (EWMA).
Position Document: Identifying Criteria for
Wound Infection. MEP Ltd, London
Daltrey D, Rhodes B, Chattwood J (1981)
Investigation into the microbial flora of
healing and non-healing decubitus ulcers. J
Clin Pathol 34: 701–5
D’Amore P (2005) Evolution of C-reactive
protein as a cardiac risk factor. Laboratory
Med 36(4): 234–8 www.medscape.com/
viewarticle/502691
Davis E (1997) Microbiology reports and
the chronic wound: a nurse’s perspective.
(Poster) European Wound Management
Association, Milan meeting 1997.
Macmillan Magazines Ltd, London
Davis E (1998) Education, microbiology
and chronic wounds. J Wound Care 7(6):
272–4
Dow G (2001) Infection in chronic
wounds. In: Krasner D, Rodeheaver
G, Sibbald R (eds.) Chronic Wound
Care: A Clinical Source Book for
Healthcare Professionals. (3rd edn) HMP
Communications, Wayne
Dow G, Browne A, Sibbald R (1999)
Infection in chronic wounds: Controversies
Key Points
8 Wound infection in covert and
overt clinical states is one of
the causes of delayed healing
in wounds healing by
secondary intention.
8 Diagnosis of wound infection
is made by identification of
clinical signs and symptoms,
which vary from patient to
patient. As clinicians differ in
their recognition of signs and
symptoms and the emphasis
they place on individual
clinical features, there can be
inconsistencies in the diagnosis
of infection.
8 Currently there is no reliable
diagnostic marker, including the
use of microbiological results,
that can be used to identify a
wound experiencing a problem
from its bioburden which has
led to inconsistency in the
prescription of antimicrobials.
8 This study explores the use of
CRP as a potential diagnostic
test for the identification
of wounds in need of
antimicrobial interventions and
for differentiation of wounds
into the different states of the
wound infection continuum
8 Mean CRP levels from patients
identified by clinical criteria
as belonging to the spreading
infection group
were significantly higher
than for patients in the other
study groups.
8 CRP levels could not
distinguish between the groups
known as colonised, critically
colonised and local infected.
Mean results from these three
groups were raised above the
normal range.
29/10/08 00:15:28
 Clinical RESEARCH/AUDIT
in diagnosis and treatment. Ostomy Wound
Manage 45(8): 23–40
Du Clos T, Mold C (2002) The role of
C-reactive protein in the resolution of
bacterial infection. Infect Dis Clin Practice
11(4): 229–33
Edmonds M, Foster A (2004) The use
of antibiotics in the diabetic foot. Am J
Surgery 187(suppl): 25S–28S
Eron L, Lipsky B, Low D, Nathwani D,
Tice A, Volturo G (2003) Managing skin
and soft tissue infections: expert panel
recommendations on key decision points.
J Antimicr Chemotherap 52(suppl S1):
i3–i17
Falanga V (2005) Wound healing and its
impairment in the diabetic foot. Lancet
366: 1736–43
Gardner S, Frantz R, Doebbeling B
(2001a) The validity of the clinical signs
and symptoms used to identify localized
chronic wound infection. Wound Repair
Regen 9: 178–86
Gardner S, Frantz R, Troia C, Eastman,
S, MacDonald M, Buresh K, Healy D
(2001b) A tool to assess clinical signs and
symptoms of localized infection in chronic
wounds: development and reliability.
Ostomy Wound Manage 47 (1): 40–7
Goodfield M (1988) C-reactive protein
levels in venous ulceration: An indication
of infection? J Am Acad Dermatol 18(5):
1048–52
Gulich M, Matschiner A, Gluck R, Zeitler
H-P (1999) Improving diagnostic accuracy
of bacterial pharyngitis by near patient
measurement of C-reactive protein (CRP).
Br J General Practice 49: 119–121
Gupta R, Singh R, Soni M (2002) C-
reactive protein (CRP) as an indicator of
sepsis in orthopaedic trauma. Ind J Med Sci
56(10): 501–7
Halbert A, Stacey M, Rohr J, Jopp-McKay A
(1992) The effect of bacterial colonization
on venous ulcer healing. Austr J Dermatol
33: 75–80
Hansson C, Hoborn J, Moller A, Swanbeck
G (1995) The microbial flora in venous leg
ulcers without clinical signs of infection.
Acta Dermatol Venereol (Stockholm) 75:
24–30
Hansson C, Jekler J, Swanbeck G (1985)
Candida albicans infections in leg ulcers
and surrounding skin after the use of
ointment impregnated stockings. Acta
Dermatol Venereol 65: 424–7
Houshian S, Zawadski A, Riegels-Nielsen P
(2000) Duration of postoperative antibiotic
therapy following revision for infected
knee and hip arthroplasties. Scand J Infect
Dis 32(6): 685–8
46 Wounds UK, 2008, Vol 4, No 4
p32-46Kingsleyok.indd 16
Howell-Jones R, Wilson M, Hill K, Howard
A, Price P, Thomas D (2005) A review
of the microbiology, antibiotic usage
and resistance in chronic skin wounds. J
Antimicrob Chemother 55(2): 143–9
Jorgensen B, Price P, Andersen KE et al
(2005) The silver-releasing foam dressing,
Contreet Foam, promotes faster healing
of critically colonised venous leg ulcers: a
randomised controlled trial. Int Wound J 2:
64–73
Kalinski C, Scnepf M, Laboy D et al (2005)
Effectiveness of a topical formulation
containing metronidazole for wound odour
and exudate control. Wounds 17 (4): 84–90
Kingsley A (2001) A proactive approach
to wound infection. Nursing Stand 15(30):
50–8
Kingsley A, Winfield-Davies S (2003) Audit
of wound swab sampling: why protocols
could improve practice. Prof Nurse 18(6):
338–43
Kingsley A, White R, Gray D, Cooper
P (2005) Using the wound infection
continuum to assess wound bioburden.
Wounds UK 1(2 suppl): 15–20
Larsson S, Thelander U, Friberg S
(1992) C–reactive protein (CRP) levels
after elective orthopaedic surgery. Clin
Orthopaed Related Res 275: 237–42
Lindholm C, Andersson H, Fossum B,
Jorbeck H (2005) Wounds scrutiny in a
Swedish hospital: prevalence, nursing care
and bacteriology, including MRSA. J Wound
Care 14(7): 313–19
Madsen S, Westh H, Danielsen S, Rosdahl
V (1996) Bacterial colonisation and healing
of venous ulcers. Acta Pathologica Microb.
Scand 104: 895–9
McGuckin M, Goldman R, Bolton L,
Salcido R (2003) The clinical relevance of
microbiology in acute and chronic wounds.
Adv Skin Wound Care 16(1): 12–23
Meaume S, Vallet D, Nguyen Morere
M, Teot L (2005) Evaluation of a silver-
releasing hydroalginate dressing in chronic
wounds with signs of local infection. J
Wound Care 14(9): 411–9
Okamura J, Miyagi J, Terada K, Hokama Y
(1990) Potential clinical applications of C-
reactive protein. J Clin Lab Anal 4: 231–5
Penhallow K (2005) A review of studies
that examine the impact of infection on the
normal wound healing process. J Wound
Care 14(3): 123–6
Pepys M (1981) C-reactive protein fifty
years on. Lancet 1: 653–7
Ribinik P, Terki M, Barraud D, et al (1997)
Pressure sores: swab results in daily
practice. In: Proceedings of the European
Wound Management Association Conference,
Milan 1997. Macmillan Magazines Ltd,
London
Schmidt K, Debus E, Jeburger St, Ziegler
U, Thiede A (2000) Bacterial population of
chronic crural ulcers: is there a difference
between the diabetic, venous and the
arterial ulcer? VASA 29: 62–70
Schraibman I (1987) Bacteriology of leg
ulcers. Phlebology 2: 265–70
Sibbald R, Williamson D, Orsted H, et al
(2000) Preparing the wound: Debridement,
bacterial balance, and moisture balance.
Ostomy Wound Manage 46: 14–35
Stahl W, Singh A, Marcus M (1985)
Responses of opsonic substances to major
trauma and sepsis. Crit Care Med 13(10):
779–82
Starr S, MacLoed T (2003) Wound
swabbing technique. Nurs Times 99(5):
57–9
Steed D, Donohoe D, Webster M, Lindsley
L (1996) Effect of extensive debridement
and treatment on the healing of diabetic
foot ulcers. J Am Coll Surg 183: 61–4
Steer J, Papini R, Wilson A, McGrouther
D, Parkhouse N (1996) Quantitative
microbiology in the management of
burn patients. 1. Correlation between
quantitative and qualitative burn wound
biopsy culture and surface alginate swab
culture. Burns 22(3): 173–6
Thomas S (1997) Assessment and
management of wound exudate. J Wound
Care 6(7): 327–30
Thomson P, Smith D (1994) What is
infection? Am J Surgery 167(1A– suppl):
7S–11S
Trengove N, Stacey M, McGechie D, Mata
S (1996) Qualitative bacteriology and leg
ulcer healing. J Wound Care 5: 277–80
Verdu Soriano J, Rueda Lopez J, Martinez
Cuervo F, Soldevilla Agreda J (2004)
Effects of an activated charcoal silver
dressing on chronic wounds with no
clinical signs of infection. J Wound Care
13(10): 419–23
White R, Cutting K (2006) Critical
colonization — the concept under scrutiny.
Ostomy Wound Manage 52(11): 50–6
Williams D, Enoch S, Miller D, Harris
K, Price P, Harding K (2005) Effect of
sharp debridement using curette on
recalcitrant nonhealing venous leg ulcers: a
concurrently controlled, prospective cohort
study. Wound Repair Regen 13(2): 131–7
Woolfrey B, Fox J, Quall C (1981) An
evaluation of burn wound quantitative
microbiology. Am Soc Clin Pathol 75(4):
532–7
29/10/08 00:15:29